Cadena de Suministro Trazabilidad 5

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 10/03/15. Copyright ASCE. For personal use only; all rights reserved.
Automated Materials Tracking and Locating: Impact Modeling and Estimation

D. Young, 1 H. Nasir, 2 S. Razavi, 3 C. Haas, 4 P. Goodrum, 5 and C. Caldas6
1
Graduate Student, Department of Civil and Environmental Engineering, University
of Waterloo, Waterloo, ON; email: d2young@engmail.uwaterloo.ca
2
Corresponding Author, Graduate Student, Department of Civil and Environmental
Engineering, University of Waterloo, Waterloo, ON; email:
hnasir@engmail.uwaterloo.ca
3
Graduate Student, Department of Civil and Environmental Engineering, University
of Waterloo, Waterloo, ON; email: snawabza@engmail.uwaterloo.ca
4
Professor, Department of Civil and Environmental Engineering, University of
Waterloo, Waterloo, ON; email: chaas@civmail.uwaterloo.ca
5
Associate Professor, Department of Civil Engineering, University of Kentucky,
Lexington, KY; email: pgoodrum@engr.uky.edu
6
Assistant Professor, Department of Civil, Architectural, and Environmental
Engineering, The University of Texas at Austin, TX; email: caldas@mail.utexas.edu
Abstract
Successful, large scale field trials were conducted on two sites in Texas and
Toronto using an integrated system of RFID tags, GPS technology, map software,
and hand held computing to automatically track materials in the projects’ respective
lay down yards. This paper addressed the unresolved research question, that how will
this technology impact projects if it is implemented upstream in the supply chain in
an integrated and automated materials management system? This question is
addressed by modeling the impact of automated materials tracking technology on
increasing visibility within the construction supply chain. It is concluded that
automated materials tracking has the potential to improve construction productivity,
cost, and schedule performance.
Introduction
The Construction industry is characterized by the fact that every project is
different from others in terms of planning, design, and execution. Likewise, the
supply networks that fuel these projects with labour, equipment, and highly
engineered materials are complex entities onto themselves. Improving the materials
management process, particularly at the site level on large construction projects, has
been identified as one area that can have a beneficial impact on the construction
supply network and the construction process as a whole (Construction Industry
Institute 1999; Kini 1999). The integration of Automated Materials Locating and

41
Copyright ASCE 2010 Construction Research Congress 2010

Construction Research Congress 2010
42 CONSTRUCTION RESEARCH CONGRESS 2010
Tracking Technologies (AMLTT), namely Radio Frequency Identification (RFID)

and the Global Positioning System (GPS), and the automation of part of the materials
management process have been proposed and shown to be a viable means of process
improvement. These technologies have thus far been shown to be well suited for
application in the construction industry, especially with regards to accurately
identifying, locating, and tracking material and equipment in a construction
environment (Song et al. 2006; Ergen et al. 2007; Razavi et al. 2008).
However, what is currently lacking from the body of knowledge, on the
application of AMLTT in construction, is an understanding of the impact that this
technology could have on a project if implemented on a larger scale. Specifically,
what impact(s) could AMLTT have if it were implemented within the construction
supply network? The objective of the research presented in this paper is to
investigate the theoretical impact that the application of AMLTT could have on the
construction materials management process if implemented within different segments
of the supply network. The specific area of the construction supply network that was
chosen for investigation was the use of AMLTT to increase the level of visibility
within the greater supply network. To examine the impact of AMLTT to increase the
level of visibility within the supply chain, the process piping supply chain was
modeled using EZStrobe, a commercially available modeling and simulation software
package.
Background/Literature Review
The success of Supply Chain Management (SCM) in the manufacturing sector has
lead to interest in its application in other industries, including the construction
industry. The application of SCM principles, such as Just-In-Time (JIT) delivery, has
seen some success in the construction industry. Tommelein (1998) was successful in
demonstrating the theoretical ability of pull-driven scheduling to improve the
performance of the pipe spool installation process by reducing on-site material
buffers, decreasing the project duration, and identifying a means in which to increase
crew productivity by altering their start date. Al-Sudairi et al (1999) simulated a steel
erection process and concluded that a moderate material buffer on site could balance
the impact of uncertainty in material delivery and productivity. Vrijhoef & Koskela
(2000) went on to examine the supply chains of three construction projects (two
residential projects and one office building). From their research it was determined
that there exist significant problems (from a Lean perspective) in a typical
construction supply chain, such as the inclusion of extra time in schedules to
counteract uncertainties and the overall nearsighted management of the supply chain
in general.
The difficulty in applying SCM to construction in all of the studies mentioned
above can be recognized to originate from the fundamental differences between
manufacturing and construction. The construction industry is dominated by ‘one-off’
projects; where ‘one-off’ refers to the lack of repetition in the construction of like
facilities (Vrijhoef & Koskela 2000). Construction schedules vary considerably
between projects, some schedules span years, while others span only a few months.
Other constraints imposed on construction projects, such as weather, site conditions,

CONSTRUCTION RESEARCH CONGRESS 2010 43
site fabrication, availability of resources (e.g. skilled and unskilled labour), and local
laws and regulations can also be highly unpredictable and variable (Koskela 1992;
Cox et al. 2006).

Materials management at the site level, in general, is focused on ensuring that the
right materials are available for work crews at the right point in time. Common
materials related issues that can arise include: lack of availability of required
material, inaccurate warehouse records, damage of critical material or equipment due
to improper storage, improperly sequenced deliveries, and large surpluses of material
at project completion (Plemmons & Bell 1995; Thomas et al. 2005). The
introduction of new technologies, which fall under the category of Automated
Materials Locating and Tracking Technology (AMLTT) and the automation of
specific aspects of the site materials management process have been proposed and
shown to be a viable means of improving this process as a whole (Song et al. 2006;
Ergen et al. 2007; Razavi et al. 2008).
Field Trials
Extensive field trials were conducted to investigate the possible impacts of
AMLTT on construction productivity. The field trials were conducted at power plant
construction projects in Rockdale, Texas and Toronto, Ontario, respectively. The
Rockdale field trial was conducted at the Sandow Steam Electric Station Unit 5
project. The second field trial was conducted at the Portlands Energy Center (PEC)
project, located in the port area of Toronto, Ontario.
RFID tags were attached to the construction materials. At the PEC, 224 pipe
spools, 22 safety valves, and 100 globe valves were initially tagged with RFID
transponders. In Rockdale, Texas, 400 structural steel components were tagged for
one boiler. On both sites, data was collected on a daily basis to update and generate
the positions of the tagged components. An individual carried a kit consisting of an
integrated GPS unit, RFID reader, and a handheld computer. As the individual moved
around the site, the GPS receiver determined its own position while the RFID reader
identified the presence of the tagged components around that position at a rate of
several thousand reads per second. The localization algorithms then processed the
collected data to provide a more accurate estimate of the location of each tagged
component. Details about these field trials and their results can be found in CII
(2008).
Formulated Simulation Models
Simulation models were formulated for investigating the increase in visibility
within the construction supply network due to the use of AMLTT. Models of the
process piping supply chain, a typical component of industrial construction projects,
were developed as the primary investigative tool for this research. Process piping,
specifically the fabrication and installation of pipe spools, was chosen as the basis of
the completed investigation due to the availability of literature which provided an
understanding of its supply chain network and due to the authors’ own field
experience with this construction activity (Howell & Ballard 1996; Tommelein 1998;

Razavi et al. 2008). Two simulation models were developed. The first simulation
model represents the existing process or situation as it typically exists in the
construction industry today utilizing a material buffer. The subsequent model(s)

represents a modified version of the existing process model in which AMLTT has
been included thereby utilizing no material buffer. The EZStrobe: General-Purpose
Simulation System (EZStrobe) was utilized to build the construction supply network
models. EZStrobe is a discrete event simulation environment that is based on the
STROBOSCOPE simulation platform.
Increasing Visibility within the Construction supply network
The series of models presented in this section were used to investigate the
premise that an increased level of information transfer within the construction supply
network, regarding the status of material and equipment, would subsequently reduce
uncertainty at the site level. As a result, potential work opportunities would be
created and taken advantage of. A work opportunity is defined as a task or series of
tasks that could potentially be started and completed earlier in the construction
schedule. These work tasks, however, are typically delayed due to uncertainty in the
ability of the supply network to provide sufficient and/ or the required material to
allow for an adequate and constant rate of production to be attained on site (i.e. the
risk of delay is considered too great). Consequently, these work tasks are put on hold
until a sufficient material buffer has been accumulated on site. For example, pipe
spool installation is not typically started until a significant portion of the total pieces
to be installed (60 to 80%) have been delivered to the site (Howell & Ballard 1996),
thus utilizing a 60 to 80% material buffer.
Existing Supply Chain Model
The entire supply network for a typical construction project is composed of a
complex system of elements. Modeling such a vast system was considered to be
beyond the scale required to complete the intended investigation as part of this
research. As such, only a single supply chain from within a greater construction
supply network was modeled using EZStrobe. The chosen supply chain was that of a
typical piping process on an industrial construction project. The piping supply chain
model presented by Tommelein (1998) proved to be an essential resource. Industrial
piping is also a segment of the construction supply network that is known to suffer
from the effects of uncertainty as described by Howell and Ballard in their 1996
report entitled Managing Uncertainty in the Piping Function. The EZStrobe model
developed to represent the existing piping supply chain was broken into two parts.
The first part of the model depicts the segments of the piping supply chain which
relate to the fabrication process. The second part of the EZStrobe model depicts the
segments which relate to on-site installation. The first part of the model is composed
of those activities which are directly related to the fabrication of pipe spools and
associated activities which are based offsite. These activities include the generation
of design specifications, the design process, the overall pipe spool fabrication
process, and the transfer of material to the site. The second part of the developed
model is composed of those activities that are directly related to the installation of the
pipe spools and other associated on site activities. The activities in this part of the

model include the material receiving process, the completion of install area
preparation, pre-installation work, and final installation. There is little to no supply
network visibility assumed to exist within this model.
Modified Supply Chain Model

AMLTT, specifically RFID, employed in conjunction with a comprehensive data
management system, offers the potential to drastically improve the ability to transfer
information within the construction supply network. This in turn influences the
potential to increase supply network visibility and subsequently decrease the level of
uncertainty held at the site level concerning the status of material and equipment yet
to be delivered. It is theorized that with decreased levels of uncertainty concerning
material and equipment deliveries that site based project managers will be able to
take advantage of work opportunities that would have otherwise been considered too
risky. Accordingly, the requirement to build up costly stockpiles of material and
equipment in project lay down yards and warehouses, prior to beginning specific
construction tasks is expected to be mitigated. This theory is explored using the
modified piping supply chain model presented in this section. In this model, however,
it is assumed that AMLTT, in the form of RFID transponders, associated readers, and
data management and communication systems, is incorporated into the piping supply
chain at the earliest possibly phase. The start of the fabrication process was
considered to be the earliest point in which AMLTT could be integrated into the
piping supply chain, as this is the first point in which a physical entity exists that a
RFID transponder can be affixed to. With the increased use of information
technology based tools such as Building Information Models (BIM), the ability to
automatically convey information regarding the status of specific items to a central
project database may also be extended into the engineering and design phase in the
future.
Figure 1 represents about a 10% sample of the complete model. We are sorry we
could not present the full model which incorporated 40 “combi’s,” 32 normals, and
dozens of probability branches and queues. These cannot be presented in this paper.
The modified EZStrobe model is divided into three parts. Part 1 focuses on the
segments of the supply chain related to the pipe spool fabrication process. Part 2, is
focused on the pipe spool installation process. Part 3, is comprised of the majority of
the additional modeling elements required to incorporate the conceptual flow of
AMLTT derived information and the mechanism by which possible work
opportunities are generated.
In the modified model, the install area preparation, pre-installation work, and
installation activities are not constrained by the quantity of loads received on site;
rather they depend on the availability of potential work opportunities that are
generated as a result of information conveyed from various stages of the supply
chain. It should be noted that a 50 day delay period to account for the completion of
site work was applied to the install area preparation activity in this model. Dynamic
based probabilities were assigned to the decision nodes associated with on site data
collection points to account for the occasions when materials arrive on-site early or
out of sequence with installation. The probability that a data signal originating from


include the design of pipe spools, the overall fabrication of pipe spools, and the
delivery of pipe spools to the site. The installation based activities include the install
area preparation, pre-installation pipe spool work, and installation activities. A

summary of the simulation results is provided in the next two sub-sections.
Summary of Existing Supply Chain Model Output

The existing model output for the case of a 60% install buffer is presented. Figure
2, shows the percent complete of each activity over time. The “Install Area Ready”
and “Pipe Ready for Install” activities commence operations 124 days after the
simulated project kick off. The last area is subsequently made ready for installation
at day 222. This finish date corresponds to a rate of production of approximately 1
area prepared every 10 days. This rate of production is in line with the assumed
overall rate for this activity. The final set of pipe spools is made ready for installation
on approximately day 226. The corresponding rate of production is therefore 3.9 pipe
spools per day. Installation is subsequently able to commence on day 131.
Installation was recorded as being complete on day 246. This translates into a rate of
production of approximately 0.9 areas per 10 days. This rate of production is slightly
less than anticipated; however, it can most likely be attributed to the accumulation of
incremental delays incurred in upstream activities. Overall, based on a combination
of 60% install buffer and 55% precedent activity buffer, the estimated optimum
schedule duration for the existing model was recorded to be 246 days. This is the
result for which the corresponding output of the modified model will be compared to.
Figure 2: Percent Complete vs. Time Figure 3: Percent Complete vs. Time
(Existing Model) Modified Model (Work Opportunities)
Summary of Modified Supply Chain Model Output

The simulation output for the modified model output is presented in the same
format as the previous model. Figure 3 provides the percent complete profiles for
each activity. Relatively constant rates of production are able to be achieved for each
of the three on-site activities under the modified model. Sufficient work opportunities
were generated to allow for the pipe spool pre-installation work activity to commence

immediately following the delivery of the first load of pipe spools on day 20. This
activity reached 100% completion 190 days later on day 210. This corresponds to an
overall rate of production of 2.1 pipe spools per day. Sufficient work opportunities
were also generated to allow the preparation of install areas to commence
immediately after the 50 day delay period to account for site work. This activity
reached 100% completion on day 139. This translates into a rate of production of 1.1
areas every 10 days. Finally, installation was able to begin on day 58 and was
subsequently completed on day 219. The resulting rate of production for this activity
was calculated to be 0.6 areas completed every 10 days.
Comparison of Schedule Performance

To compare the performance of the modified model to that of the existing model
in terms of schedule, three parameters were examined for each site based activity.
The parameters examined include the start and end time of each activity and the
ensuing rate of production. Each activity is compared separately in the following
sub-sections.
Comparing the Rate of Install Area Preparation

The corresponding percent complete profiles for the install area preparation
activity for both the existing and modified models from the previous sections are
displayed together in Figure 4 (a) below. The transparent overlays correspond to the
resource allocation profiles for each instance of this activity. The elimination of the
material buffer and the subsequent change over to a reliance on work opportunities
generated via AMLTT has the effect of moving ahead the start time of the install area
preparation activity. The resulting difference between the start times is 73 days in
favor of the modified model. The overall length of time to complete the preparation
of all 10 install areas was also reduced from 99 to 89 days; however, this reduction
can be attributed to randomness in the simulation engine and would not be expected
to occur on average. The subsequent rate of production of each instance of this
activity is relatively the same. Accordingly, the resource allocation profile is the
same for each case at 1 full crew for 100% of the activity duration.
Comparing the Rate of Pre-Install Pipe Work

The percent complete profiles for the pre-install pipe work activity for both the
existing and modified models are displayed together in Figure 4 (b). The resource
allocation profiles for each instance of this activity are overlaid in the corresponding
transparent colour. The effect of eliminating the reliance on material buffers as a
primary constraint on operations is more evident when comparing the percent
complete profile for the pre-install pipe work activity for each model. In the case of
the existing model the pre-install work is completed over a relatively short period of
time of 102 days as previously reported. In the case of the modified model, the start
time of the pre-install work is able to be pushed forward by just over 100 days. The
overall duration, however, is increased by a factor of two. This is caused in part due
to increased crew idle time, which is manifested in the form of disjoints in the percent
complete profile for the modified model. Accordingly, the rate of production
between the existing and modified model is decreased by a factor of 2. However, this

can be compensated for by reducing the crew size by a half as indicated by the
corresponding resource allocation profile. While the overall activity duration is
increased, most project managers should appreciate the greater flexibility that this
alternative provides and improved productivity.
Figure 4 (a) Figure 4 (b) Figure 4 (c)

Figure 4: Comparing Rates of Productivity – (a) Install Area Preparation
Activity; (b): Pre-Install Pipe Work Activity; (c): Installation Activity
Comparing the Rate of Installation

The percent complete profiles for the installation activity for both the existing and
modified models are displayed together in Figure 4 (c). The resource allocation
profiles for each instance of this activity are overlaid in the corresponding transparent
colour. Similar differences in the percent complete profiles for the installation activity
between the existing and modified models can be seen here as compared to the
previous activity. The elimination of the material buffer allows for the start time of
the activity to be shifted forward, in this case by 77 days. The overall activity
duration under the modified model is increased by a factor of 1.4 from 115 to 161
days, which in turn decreases the overall rate of production. However, like the
previous case, this can be compensated for by reducing the overall crew size by
slightly less than 30%.
Conclusions
Based on the completed comparison it is evident that the integration of AMLTT
into the construction supply chain has the potential to improve or at least maintain
construction schedule performance, while reducing the dependency on costly material
stockpiles. It is recognized that under the AMLTT based model the rate of
production of two thirds of the site based activities was reduced significantly and
crew idle time increased. However, as was shown, this can be compensated for by
reducing the applicable crew size. The ability to start work tasks earlier in the
construction schedule as a result of an increased state of knowledge of the supply
situation should also provide project managers with a greater flexibility to account for
other unexpected events that may hinder project performance. In addition, the overall
duration of the theoretical piping operation was reduced from 246 to 219 days in
favor of the AMLTT based model; a reduction in schedule duration of approximately

11%. An improvement in schedule performance of even 2 to 4% would translate into

significant cost savings on any typical industrial construction project. While
completely hypothetical, the results presented here serve to demonstrate the

beneficial impact on schedule performance that could potentially be achieved via the
integration of AMLTT within a single construction supply chain such as piping, let
alone the greater construction supply network as a whole.
References
Al-Sudairi, A. A., Diekmann, J. E., Songer, A. D., & Brown, H. M. (1999).
Simulation of Construction Processes: Traditional Practices Versus Lean
Principles. IGLC-7 (pp. 39-50). Berkeley, CA: International Group for Lean
Construction.
Construction Industry Institute. (1999). Procurement and Materials Management: A
Guide to Effective Project Execution. University of Texas at Austin. Austin,
Texas: Construction Industry Institue.
Construction Industry Institute. (2008). RT-240 Leveraging Technology to Improve
Construction Productivity,Publication No. 240-11, University of Texas at Austin.
Austin, Texas: Construction Industry Institute.
Cox, A., Ireland, P., & Townsend, M. (2006). Managing in Construction Supply
Chains and Markets. London, United Kingdom: Thomas Telford Publishing.
Ergen, E., Akinci, B., & Sacks, R. (2007). Tracking and locating components in a
precast storage yard utilizing radio frequency identification technology and GPS.
Automation in Construction , 16 (3), 354-367.
Howell, G. A., & Ballard, H. G. (1996). RR 47-13 Managing Uncertainty in the
Piping Function. The University of Texas at Austin. Austin, Texas: Construction
Industry Institute.
Kini, D. (1999). Materials Management: The Key to Successful Project Management.
Journal of Managment in Engineering , 30-34.
Koskela, L. (1992). Application of the New Production Philosophy to Construction.
Stanford: Stanford University.
Plemmons, J. K., & Bell, L. C. (1995). Measuring Effectiveness of Materials
Management Process. Journal of Management in Engineering , 11, 26-32.
Razavi, S. N., Young, D. A., Nasir, H., Haas, C. T., Caldas, C., Goodrum, P., et al.
(2008). Field Trial of Automated Material Tracking in Construction. CSCE 2008
Annual Conference. Quebec City.
Song, J., Haas, C. T., & Caldas, C. (2006). Tracking the Location of Materials on
Construction Job Sites. Journal of Construction Engineering and Management ,
132 (9), 911-918.
Thomas, H. R., Riley, D. R., & Messner, J. I. (2005). Fundamental Principles of Site
Material Management. Journal of Construction Engineering and Management ,
131 (7), 808-815.
Tommelein, I. D. (1998). Pull-Driven Scheduling for Pipe Spool Installation:
Simulation of Lean Construction Technique. Journal of Construction Engineering
and Managment , 124 (4), 279-288.
Vrijhoef, R., & Koskela, L. (2000). The Four Roles of Supply Chain Management in
Construction. European Journal of Purchasing & Supply Management , 6, 169-
178.


Cadena de Suministro Trazabilidad 5

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cadena de Suministro Trazabilidad 5

Uploaded by

Copyright:

Available Formats

Downloaded from ascelibrary.org by UNIVERSITE LAVAL on 10/03/15. Copyright ASCE. For personal use only; all rights reserved.

Automated Materials Tracking and Locating: Impact Modeling and Estimation

Copyright ASCE 2010 Construction Research Congress 2010

Tracking Technologies (AMLTT), namely Radio Frequency Identification (RFID)

Copyright ASCE 2010 Construction Research Congress 2010

Cox et al. 2006).

Copyright ASCE 2010 Construction Research Congress 2010

construction industry today utilizing a material buffer. The subsequent model(s)

Copyright ASCE 2010 Construction Research Congress 2010

network visibility assumed to exist within this model.

Modified Supply Chain Model

Copyright ASCE 2010 Construction Research Congress 2010

Copyright ASCE 2010 Construction Research Congress 2010

area preparation, pre-installation pipe spool work, and installation activities. A

Summary of Existing Supply Chain Model Output

Summary of Modified Supply Chain Model Output

Copyright ASCE 2010 Construction Research Congress 2010

Comparison of Schedule Performance

Comparing the Rate of Install Area Preparation

Comparing the Rate of Pre-Install Pipe Work

Copyright ASCE 2010 Construction Research Congress 2010

Figure 4 (a) Figure 4 (b) Figure 4 (c)

Comparing the Rate of Installation

Copyright ASCE 2010 Construction Research Congress 2010

11%. An improvement in schedule performance of even 2 to 4% would translate into

completely hypothetical, the results presented here serve to demonstrate the

Copyright ASCE 2010 Construction Research Congress 2010

You might also like